The present invention relates to the field of speech user interfaces.
Applications which utilize a speech interface operate via speech recognition to identify a speech (e.g., spoken) command. Spoken commands must utilize one or more correct phrases in order to actuate a function provided by the application. For example, to close a vehicle window within a speech enabled automobile, phrases such as “Raise the window” or “Roll-up the window” can be equivalent, resulting in the closing of the window. In almost all cases, however, speech commands and/or phrases must be discovered by trial and error, by referring to a list provided, or in an off-line manual (e.g., vehicle manual).
These solutions fail to fully utilize the speech interface functionality to address the discovery of commands. Provided lists can be organized into task domains, which do not permit easy identification of a specific command readily. For instance, to manually discover a tune station speech command, a user must select “Radio Control” task domain, select “Mode Management”, then select “Tuning” to discover the appropriate speech command. Further, many functions can be associated with a single speech command. Consequently, identifying an appropriate speech command to utilize can be troublesome in many scenarios. For example, within an automobile, traditional manual speech command discovery can result in safety issues by distracting a driver from maintaining road awareness.